Our Journey to a Software Testing Ice-Cream Cone approach

30-minute New Voice Talk

A Software Testing Ice-Cream Cone approach can lead to benefits, challenges, and (well-known) pains.

Timetable

11:55 a.m. – 12:25 p.m. Thursday 15th

Room

Room F1 - Track 1: Talks

Audience

Testers, Team leaders, and project managers.

Key-Learning

  • 'Benefits' of our Software Testing Ice-Cream Cone experience
  • 'Pains' of our Software Testing Ice-Cream Cone experience
  • Challenges of our Software Testing Ice-Cream Cone experience

Three years ago, a client from a Brazilian private institute decided to make some changes on the software development process for its four agile projects with the institute. At that time, UXs designers have the user stories and its acceptance criteria on a spreadsheet, QA manual has its scenarios in a specific tool, QA automation has end to end tests coded in Java with Selenium, and devs have their unit tests coded in PHP. As can be noticed, each sub-team uses the UX artifact (spreadsheet) as a foundation for their artifacts. In the hope that the team will have a single source of truth for the project’s requirements, our client decided to adopt a Business Readable - Domain Specific Language called Gherkin.

At the beginner, all sub-teams were open-minded to the new process, and everything will start with UX writing the .feature in a natural language. However, after two sprints (?), the UX team decided to go back to the old style arguing that we have a lot of changes on the ACs, and it was impracticable to them continue with this approach. On the other hand, our QA team decided to follow the new approach, so during the development of the stories, QA writes a .feature using a pre-defined dictionary and after the story was released, we could automate it. The origin idea was that the dev team could use that artifact as an input to write unit/integration tests. Besides, we thought that the tests could be helpful as documentation because they were written in English so that everybody could understand it.

After some sprints, we finally had something to tell our clients and community about our experience with that approach. In reality, as expected for a well-known anti-pattern approach, not everything was a bed of roses. During that period, we could notice some good reasons for adopting this "new" approach, but we were also feeling some pains, and going through some challenges. Given these points, we would like to share that experience with the community, and it can be helpful to someone who is going through a similar process and context.

Related Sessions

There are currently no related sessions listed.